An attempt to create an efficient non-invasive blood pressure monitor, using a simple circuit prototyped with an Intel Edison. The hardware includes a Light Emitting Diode (LED) for measuring the amount of blood, and for pulse oximetry (non-invasive method for measuring a patient’s Oxygen saturation). The sensors take raw data input, and process them using a filter algorithm similar to a low pass filter (LPF), and averages the data. The device is capable of filtering out ambient light using a calibration algorithm. Once the device is calibrated with the current lighting conditions, it can read the person's blood pressure (using pulse oximetry) by simply checking the amout of light that passes through their fingers. This data is not stable, and needs to be filtered out (we ended up using a custom made algorithm inspired by the LPF).

The inspiration for this project was diabetic patients who have to check their blood glucose multiple times a day. This process of checking for blood glucose often includes pricking their fingers (which can be quite painful). There are multiple papers published online for example, IEEE NIRLED). The problem with non-invasive blood glucose monitoring, however, is that it is not very efficient. We wanted to test this for ourselves, and hence, we decided to create a simpler version of the blood glucose monitor, based off of the dozens of articles published online. With our tests, we had a 70% success rate under appropriate lighting conditions. In order to confirm our results, we had our teammates run down the stairs and back up again, and tested our blood pressure after that. The recordings were exactly as expected. The only situation in which it seemed to fail, was under very high lighting conditions, when the ambient light is much more powerful than the LED light we were using (we used a standard LED commonly found with arduino kits).

Share this project: